Distributed LQR-based observer design for large-scale multi-agent networks (2001.05760v1)

Abstract: In this paper, network of agents with identical dynamics is considered. The agents are assumed to be fed by self and neighboring output measurements, while the states are not available for measuring. Viewing distributed estimation as dual to the distributed LQR problem, a distributed observer is proposed by exploiting two complementary distributed LQR methods. The first consists of a bottom-up approach in which optimal interactions between self-stabilizing agents are defined so as to minimize an upper bound of the global LQR criterion. In the second (top-down) approach, the centralized optimal LQR controller is approximated by a distributed control scheme whose stability is guaranteed by the stability margins of LQR control. In this paper, distributed observer which minimizes an upper bound of a deterministic performance criterion, is proposed by solving a dual LQR problem using bottom-up approach. The cost function is defined by considering minimum-energy estimation theory where the weighting matrices have deterministic interpretation. The presented results are useful for designing optimal or near-optimal distributed control/estimation schemes.